Estimating the Effects of Family Group Decision Making when Responses to Treatment Vary Among Observationally Identical Families


Previous evaluations of Family Group Decision Making suggest participation does not affect the likelihood of an out-of-home placement or a re-referral; however, many of these studies only consider the impact of Family Group Decision Making on families receiving in-home services. To capture the interdependence between re-referrals and out-of-home placements, I estimate a latent-variable model that allows for selection on out-of-home placement in the re-referral equation. I assume the unobservables in the outcome and participation decision equations observe a normal factor structure, and I calculate various mean treatment parameters where responses to treatment vary among observationally-identical families.